Image Processing Method for Movement Detection and Compensation

1. An image processing method for improving image quality, the method comprising: selecting one of a plurality of two-dimensional pixel matrix source images of an imaged target as a reference image; searching for common points in the source images which correspond to each other; detecting movement of the imaged target in the source images by forming a movement estimate for a pixel in a section of at least one source image by weighting a movement vector of the pixel on a basis of a distance from the common points in the section and a recoding moment of the section; compensating the detected movement between the reference image and the at least one source image differently in at least two different sections of the at least one source image on a basis of an order of recording moments of the at least two different sections; and combining the reference image and the detected movement compensated at least one source image for forming an improved image.

2. The method as claimed in claim 1 , further comprising: estimating a reliability of the common points by determining a similarity between the common points in different source images.

3. The method as claimed in claim 1 , further comprising: estimating a reliability of the common points by determining a movement of each of the common points in a direction from the reference image to an image of the at least one source image and in an opposite direction from the image of the at least one source image to the reference image, and accepting the common points if the determined movements correspond to each other within a threshold range, and rejecting the common points if the determined movements deviate from each other more than the threshold range.

4. The method as claimed in claim 1 , further comprising: estimating movements for pixels in the at least two different sections by a weighted least square algorithm or an iterative M-estimator.

5. The method as claimed in claim 1 , further comprising: searching for a random movement area in the source images where a distribution of intensities of pixels is random and a variation of the distribution is above a preset threshold; and setting the random movement area from only the reference image to the improved image, if the random movement area is found.

6. The method as claimed in claim 1 , further comprising: recording each of the source images by scanning over a pixel matrix of a detector such that each section is recorded at a different moment.

7. The method as claimed in claim 1 , further comprising: forming an intensity of a pixel of the improved image on a basis of intensities of neighbouring pixels which surround the pixel three-dimensionally in the reference image and in the at least one source image.

8. The method as claimed in claim 7 , wherein an intensity of a pixel is or intensities of a group of pixels are based on similarities of local pixel structures in the at least one source image.

9. An apparatus for improving quality of an image, the apparatus comprising at least one processor programmed to: select one of a plurality of two-dimensional pixel matrix source images of an imaged target as a reference image; search for common points in the source images which correspond to each other; detect movement of the imaged target in the source images by forming a movement estimate for a pixel in a section of at least one source image by weighting a movement vector of the pixel on a basis of a distance from the common points in the section and a recoding moment of the section; compensate the detected movement between the reference image and the at least one source image differently in at least two different sections of the at least one source image on a basis of an order of recording moments of the at least two different sections; and combine the reference image and the detected movement compensated at least one source image for forming an improved image.

10. The apparatus as claimed in claim 9 , further comprising: one or more memories including computer program code; wherein the one or more memories and the computer program code being configured to, with the at least one processor, cause the apparatus at least to perform the selection of the reference image, the detection of movement, the compensation of the detected movement and the combination of the reference image and the at least one source image.